SpaceX Posts Strong Q2 Numbers, Then Watches the Stock Slide Anyway
The numbers came in hot. SpaceX reported second-quarter revenue of $7.8 billion, nearly doubling year-over-year and clearing analyst expectations of $6.9 billion by a wide margin. Adjusted EBITDA tripled. Every business line — rockets, Starlink connectivity, and AI — beat forecasts. On paper, this was the kind of debut earnings report a newly public company dreams about.
The market did not care.
Shares fell more than 5% when results dropped Tuesday evening and continued sliding 6% to 8% in after-hours trading. The culprit was capital expenditure. SpaceX reported $18.4 billion in capex for the quarter, with nearly $16 billion directed at AI compute infrastructure. Analysts had modeled $13.2 billion. The first quarter came in at $10.1 billion. In a single quarter, spending accelerated by more than $8 billion.
'The stock is down because the capex for the AI segment was more than double what was expected,' said Melissa Otto, global head of Visible Alpha research at S&P Global.
Starlink Is the Story — If You Give It Time
CEO Elon Musk used the conference call to make the bull case, loudly. Starlink posted $4.3 billion in revenue for the quarter, up 66% year-over-year. The subscriber base doubled to 12 million, with 1.7 million new customers added in Q2 alone. Airlines are signing on: American, Southwest, Virgin Atlantic, Iberia, and Aer Lingus all inked Starlink agreements during the period.
'I think people are really underestimating Starlink,' Musk told analysts, adding that it is 'not out of the question' that Starlink eventually delivers a majority of the world's internet. He put that timeline at 'less than 10 years.'
The AI segment grew revenue 247% to $2.6 billion, driven by cloud hosting agreements with Google and Anthropic, Grok subscriptions, and X advertising. SpaceX also reported $6.7 billion in cloud services deals already signed for the second half of the year.
The $1 Trillion Target Moves Closer — On Paper
Musk declared that SpaceX's internal target for $1 trillion in annual revenue has moved forward a full year, from 2031 to 2030, with what he called 'a non-zero chance' of hitting the mark in 2029. Otto of Visible Alpha called the figure ambitious but not impossible: 'Given the pace of revenue growth and investment this quarter, if the company remains in acceleration mode, it's possible.'
Musk also floated plans for moon-based robots to scale manufacturing and build a mass accelerator with solar production — even acknowledging it sounded 'totally nuts' — before pivoting back to the near-term case.
A further headwind looms Thursday, when close to a billion insider shares become eligible for sale for the first time, a lock-up expiration that historically adds downward pressure on newly public stocks. SpaceX hit a $2 trillion valuation at its June IPO and has since shed close to $500 billion in market cap from that peak.
CEO Times Take
The SpaceX story is a clean test of a principle the market has been re-learning all year: revenue growth without a credible path to capital efficiency is not enough. Investors rewarded Microsoft and Amazon this quarter for pairing accelerating revenue with blowout backlog numbers; they punished Alphabet and Tesla — and now SpaceX — for letting capex run ahead of the return timeline. That is not a verdict on the long-term thesis. Starlink's subscriber growth, the AI pipeline, and the airline partnerships are real. But free enterprise demands a return on capital, not just a vision of moon factories. Until SpaceX can show the spending curve bending toward margin, the market will keep asking the same question Musk has not yet answered with numbers: when does the investment pay?
